Example plot of what would be nice to show in the lensing chapter. Of course, we expect to coordinate with the inflation forecasters to make sure plot is consistent with what is there.
Specs used are below:
10% of sky
1uk-arcmin noise in pol for foreground cleaned map
lmin = 40 and lmax=4000
EB lensing only
The errors are:
sigma(r) = 0.00094 ~ 1e-3 without delensing
sigma(r) = 0.00018 ~ 2e-4 with delensing
which is about a factor of 5 improvement due to delensing. The second and third plots show that about 80% of the Cl_BB power is removed with delensing at e.g. ell=200.
This factor of ~5 improvement seems consistent with Figure 3 in Smith et al. (1010.0048), which seems to have about a factor of 6 improvement for a 2' beam.
If we increase the noise to 2uk-arcmin in pol, then I think we match the case Victor showed in telecon better (http://users.physics.harvard.edu/~buza/20160531_fisher/) in that we agree with the factor of 2.5 improvement from delensing shown in that Table 4. However, our sigma(r) errors are a factor of 2 better, maybe due to the different experimental setups.